Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
You may not further distribute the material or use it for any profit-making activity or commercial gain
You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from orbit.dtu.dk on: Apr 19, 2022
Modeling pesticides emissions for Grapevine LCA: adaptation of Pest-LCI model toviticulture
Christel, Renaud-Gentié ; Dijkman, Teunis Johannes; Bjørn, Anders; Birkved, Morten
Published in:Proceedings of the 9th International Conference on Life Cycle Assessment in the Agri-Food Sector
Publication date:2014
Document VersionPublisher's PDF, also known as Version of record
Link back to DTU Orbit
Citation (APA):Christel, R-G., Dijkman, T. J., Bjørn, A., & Birkved, M. (2014). Modeling pesticides emissions for Grapevine LCA:adaptation of Pest-LCI model to viticulture. In R. Schenck, & D. Huizenga (Eds.), Proceedings of the 9thInternational Conference on Life Cycle Assessment in the Agri-Food Sector (pp. 1075-1085). ACLCA.
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
Modeling pesticides emissions for Grapevine Life Cycle Assessment:
adaptation of Pest-LCI model to viticulture
Christel Renaud-Gentié1,*, Teunis J. Dijkman2, Anders Bjorn2, Morten Birkved2
1 UPSP GRAPPE, UMT VINITERA, SFR Quasav, Groupe ESA,55, rue Rabelais, BP30748, F-49007 Angers, Cedex 01, France 2 Technical University of Denmark, DTU Management Engineering, Produktionstorvet Building 424, DK - 2800 Kgs. Lyngby, Denmark
*Corresponding author. E-mail: [email protected]
ABSTRACT
This paper presents a tailored version of PestLCI 2.0, the most advanced life cycle inventory model for quantification of organic pes-
ticide emissions from arable land, customized to appropriately account for viticulture specificities affecting pesticides emissions from
vineyards. PestLCI 2.0 customization is further supported by the calculation of USEtox™ freshwater ecotoxicity characterization factors
for active ingredients relevant in viticulture. Case studies on two different vineyard management systems illustrate PestLCI 2.0 model ap-
plication.
The customization of the PestLCI 2.0 model includes addition of 29 active substances, 9 application techniques, interception by a dual
canopy (vine /grass cover), new soil and climate databases and further account of multiple vinerow treatment.
Four substances dominate the overall toxicity profiles. Comparing the results obtained with PestLCI 2.0 with existing static emission
quantification approaches results reveals that PestLCI 2.0 yields considerable lower emission loads and consequently, lower toxicity im-
pact burdens. The issue of accounting for inorganic substances is discussed.
Keywords: plant protection product, LCA, vineyard, modelling, emissions, fate
1. Introduction
Wine production and hence viticulture are an important economic and cultural sector in main parts of the
world including Europe. Viticulture is also a major pesticide consuming sector with app. 20 kg formulated pesti-
cides/ha/year (of which app. 15 kg/ha are elemental sulfur) within the European Union (Endure 2010).
Pesticides emissions are therefore a crucial topic to be addressed when performing wine and/or grape produc-
tion related life cycle assessments (LCAs). Due to the lack of specific inventory models suited for pesticide
emission quantification, most of the published wine LCA studies either neglect pesticide emissions (Point et al.
2012; Pattara et al. 2012; Bosco et al. 2011; Ardente et al. 2006), simply assume that 100% is emitted to the soil
(in accordance with the default Ecoinvent approach applied for pesticides (Nemecek and Schnetzer 2011)) or
propose a fixed partition as (Neto et al. 2012) with 75 % of pesticides emitted to soil and 25% to the air.
PestLCI (Dijkman et al. 2012; Birkved and Hauschild 2006) is currently the most advanced LCI model for
pesticides emissions quantification from arable land (van Zelm et al. 2014). Despite the fact that the model has
been applied in wine/grape production LCAs (Villanueva-Rey et al. 2014; Vázquez-Rowe et al. 2012), PestLCI
2.0 still does not take into account certain viticulture specificities such as double cropping systems, vertical ap-
plication techniques etc. which differentiate viticulture from other crops and influence the emission patterns of
pesticides. Characterization of the impacts from pesticide related emissions specific to viticulture and fruits or
vegetable crop production furthermore requires characterization factors for these chemical compounds.
This paper presents a PestLCI 2.0 version customized to appropriately account for viticulture specificities af-
fecting the emission quantities and patterns of pesticides from vineyards. The application of the customized in-
ventory model is illustrated by calculation of potential ecotoxicity impacts through combination of emission
quantities and freshwater ecotoxicity characterization factors calculated using the USEtox™ characterization
model. The combination of the PestLCI 2.0 and USEtox™ models is illustrated through two case studies on
vineyard technical management routes1 (Renaud-Gentié et al. 2014; Renaud-Gentié et al. 2013).
1 technical management routes (TMRs): logical successions of technical options designed by the farmers (Renaud-Gentié et al. 2014))
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
2. Methods
2.1. Pest LCI
PestLCI is a dedicated inventory model intended to provide emission quantities for use in relation to the char-
acterization of chemically induced impacts caused by emissions of pesticides from arable land/technosphere to
the environment/ecosphere. The most recent version of the model, version 2.0, is described in Dijkman et al.
(2012). This version of the model covers app. 90 active ingredients of various types of pesticides, 25 European
climate profiles and 7 European soil profiles.
Despite the rather extensive coverage in terms of pesticides, climates and soils, the most recent model version
was developed with conventional agriculture in mind and viticulture was hence not addressed specifically. In or-
der to improve the viticulture specificity of the model, PestLCI version 2.0 was updated with 28 pesticides fre-
quently used in European viticulture, 5 French soil profiles, 22 French temperate maritime climate profiles, 34
crop stages and cover crop combinations + 13 direct spraying situations as well as 9 pesticide application tech-
niques typically applied in European viticulture.
2.2. Identification of specific needs for pesticides emissions modeling in viticulture
2.2.1. Pest and weed management in viticulture
The main pests damaging the vine canopy are primarily the fungi downy mildew (Plasmopara Viticola) and
powdery mildew (Uncinula Necator), which necessitate fungicide treatments. The other fungi and the main in-
sect pests (moths, leafhoppers and phytophagous mites) are not systematically treated. Vineyard management in-
cludes also weed control, since weed presence can affect vine growth by competition for water and nutrients.
Most of these pests require specific pesticide active ingredient substances (ai.s). The risk of resistance acquisi-
tion by the pests implies frequent change of ai.s, in conventional2 viticulture especially for ai.s presenting a sin-
gle-site action. Vineyard treatment programs therefore usually involve a variety of pesticide ai.s. This is however
not the case is in organic viticulture, where the fungicides used (primarily copper and/or sulfur based fungicides)
have multi-site action, while weeds in organic viticulture are mechanically controlled.
2.2.2. Active substances
Vineyard management involves, in French context, on average 14 to 16 pesticide applications per year at full
recommended application rates (Mézière et al. 2009; Ambiaud 2012b) with a high interregional and inter-annual
variability mainly related to climatic conditions and winegrowers practices. Organic and inorganic or partially
inorganic pesticides ai.s are used by viticulture.
Conventional pest management in viticulture relies on generic farming organic pesticides ai.s, but also on
more crop specific organic pesticide ai.s typically used on vegetables or in orchards. All these crop specific ac-
tive substances were not included in PestLCI 2.0 version presented by (Dijkman et al. 2012). The substances
used in the cases studied in the project at hand were listed. Compilation of data on the properties of these sub-
stances used in viticulture, however missing in the PestLCI 2.0, was conducted applying specialized chemical
databases: “e-phy” (MAAF and ONPV 2013) for correspondence between commercial name and active sub-
stance, “PPDB” (University-of-Hertfordshire 2013), “Toxnet” (US-National-Library-of-Medicine 2013) as well
as “Chemspider” (Royal-Society-of-Chemistry 2013) for main chemical and physical characteristics. Data gaps
were compensated for applying the Quantitative Structure-Activity Relationships included in the EPI SuiteTM
(US-Environmental-protection-Agency 2012).
However, conventional and organic viticultures also apply inorganic sulfur (S8) and copper-based fungicide
ai.s. to manage powdery mildew (mainly sulfur) and downy mildew (mainly copper). They also use other inor-
ganic ai.s like Ammonium thiocyanate (herbicide) or partially inorganic like Fosetyl-Al (fungicide). However,
PestLCI 2.0 is designed only for modelling of organic pesticide a.i. emissions. For this reason, inorganic pesti-
cides were not included in this study.
2 « Conventional » is used in this paper to designate non-organic plant protection practices
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
Some pesticide ai.s applied in the studied vineyards, especially organic vineyards are derived from micro-
organisms like Spinosyn, or plants like Pyrethrum. The physical-chemical and fate properties of these pesticide
ai.s were found in Bio-Pesticide DataBase (BPDB) (University-of-Hertfordshire 2012) and some of the previously
cited databases. A second fraction of “others pesticides and pesticide ingredients” are substances like algae ex-
tracts and pesticide formulation additives which are not considered in this study, due to lack of information about
the nature and quantity of these substances. Formulation additives may however despite of their exclusion of this
study, contribute considerable to toxicity of the pesticide formulation (Brausch and Smith 2007).
2.2.3. Spraying equipment
Vertical shoot positioning is the most frequent training system3 for vineyards in France and is found in many
other wine producing countries. The case studies assessed in the paper at hand only cover vertical shoot position-
ing trained vineyards.
The types of spraying equipment involved in viticulture are multiple, which makes the task of modeling their
individual characteristics a challenge. Herbicides are most often applied using specific sheltered booms to avoid
herbicide drift and hence deposition on vine leaves (which would damage/kill the wines). We chose to model
these application techniques as “soil incorporation” in Pest LCI since they induce very low drift. The sprayers
designed for canopy and grapes spraying can use different modes of droplets production:
- non air-assisted spray, the droplets are formed by the acceleration of the liquid under pressure in the
nozzle, and applied to the leaves by the pressure,
- airblast, the droplets are formed through nozzles, and are then subsequently transported to the leaves by
an air flux which further shakes the wine leaves, thereby facilitating the penetration of the substance into
the canopy,
- pneumatic where air accelerated to app. 300km/h meets the liquid containing the pesticide formulation
and fragments the solution into very fine droplets and conveys it to the leaves.
Different shapes of the ventilators and of the sprayers themselves lead to different patterns in terms of spray-
ing quality and drift generation. These different shapes also permit either a direct spraying of each face of the
canopy or a spraying from the top of the rows.
A last type of sprayer was designed for drift reduction: the tunnel sprayer where specific panels prevent from
drift and collect the non-intercepted spray mixture and recycle it for re-use.
PestLCI 2.0 takes into account the type of sprayer in order to quantify the drift calculations through drift
curves. Sprayers representing the above presented application techniques were however not available in PestLCI
2.0. In the present customization, 9 sprayers with different modes of action were included. For the customization
and the viticulture specific application techniques we used data from (Codis et al. 2011) who conducted drift
tests according to International Organization for Standardization (ISO) protocol (ISO 2005), on different vine-
yard spraying equipment types in France. The sprayers successfully tested by (Codis et al. 2011) plus a tunnel
sprayer tested by (Ganzelmeier 2000) were included in the viticulture customized PestLCI 2.0 version. Defining
within this list of the 9 sprayers, the ones that are the more similar to the ones we have in our case studies was
done through discussion with the author S. Codis (Codis 2014).
Since the aforementioned drift tests used to derive spray drift curves were conducted on vines with one leaf
area, PestLCI 2.0 was adapted in order to correct for spraying on vines with different leaf areas. Modelling of
custom spray techniques covering various adaptations of existing spraying equipment is considered beyond the
scope of this paper.
According to the type of sprayer, the winegrower can choose to spray 1 to 4 rows of vines simultaneously.
The number of rows treated plays a role in wind drift calculation in PestLCI 2.0. It has been taken into account
using the parameter” nozzle distance on the model”, the distance entered being the actual width (in meter) treat-
ed at the same time.
3 Training system : type of trellis and shoot positioning resulting to a given shape of the vine canopy and position of grapes.
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
2.2.4. Accounting for primary distribution in double cropping system
The primary distribution process is defined in PestLCI by 3 factors: wind drift (fd), pesticide deposition on
soil (fs) and pesticide deposition on leaves (fl) (Birkved and Hauschild 2006). The two latter are based on
(Linders et al. 2000) interception factors for single crops at different growing stages.
Concerning interception by vine canopy, PestLCI 2.0 includes interception values for vine at four different
development stages I, II, III, and IV based on (Linders et al. 2000). A stage 0 has been added to take into account
situations of leafless vines. The interception factor has been estimated from the orchard dormancy stage intercep-
tion factor (0.2) given by (Linders et al. 2000) with a division by 2 due to the difference of perennial parts im-
portance between fruit trees and vines. The value of fl for vine 0 was therefore set to 0.1.
On-field measurements of spraying mixture deposition and losses on vineyards were made by Sinfort (Sinfort
2014) (Sinfort et al. 2009) and on artificial vineyard by (Codis 2014). Distribution ratios of spray mixture be-
tween vine canopy, soil and air at 2.5 m above the soil were obtained by these authors in vineyard conditions
similar to the ones we study (rows width, types of sprayers). The fractions going to air during spraying measured
by (Codis 2014) and (Sinfort et al. 2009) were considered for introduction in PestLCI 2.0 as being i)partly con-
veyed by wind drift out of the parcel (i.e. advective transport), and ii)partly falling back on vegetation and bare
soil of the parcel (i.e. sedimentation). This is because no quantification of direct volatilization is possible (Jensen
and Olesen 2014) due to the complexity of drivers combination (properties of the spray liquid, drops size and
drops surrounding conditions)(Gil et al. 2007) and the lack of available data for some of them. As PestLCI 2.0
calculates the quantity of drifted pesticide on the basis of the dose applied and before calculating leaf intercep-
tion, we decided to apply a drift quantity correction ratio based on the pesticide fraction going to air. Full vegeta-
tion (stage III) was given the 1:1 ratio because sprayers drift curves were established on that stage.
For accounting of direct spraying with hand hose in early stages or young vineyards, and direct spraying on
grapes, 13 specific interception factors were defined through expertise.
Cover cropping on vineyard soil is a developing management scheme with nearly half of French vineyard
temporarily or permanently covered (Ambiaud 2012a). It offers multiple advantages, provided the competition
for water and nitrogen isn’t excessive for the vineyard (Celette et al. 2009). This second canopy under the vine-
yard (e.g. spontaneous species, oats, clover or fescue) contributes to pesticide interception (primary distribution)
and fate (secondary distribution). The interception by the cover crop varies according to the width of the cover
crop strips estimated in percentage of the width of the vine inter-row, and according to cover crop canopy densi-
ty.
In the case of mixed cropping (vine + cover crop), a complementary interception factor (f) needs to be added
for cover crop. The structure of PestLCI 2.0 being fixed with 3 f entries, a combined f has been calculated from
fvine and fcovercrop. Based on calculation routines, 34 combined interception factors are now available in the custom-
ized PestLCI 2.0 model covering the most current situations spanning possible combinations of vine develop-
ment stages, cover-crop strip width and cover crop canopy density. Examples of such combinations are given in
Table 1.
Table 1: Examples of combined interception factors for vine/cover crop mixed cover (f = interception factor)
Stage cover density
% of soil
surface cov-
ered by
grass fvine fcovercrop % spray
lost in air
% intercepted
by vegetal soil
cover (calcula-
tion) f global
0 none 0 0.1 0.3 30% 0% 0.10
II weak (30%) 100% 0.5 0.3 30% 6% 0.56
II high (70%) 80% 0.5 0.7 30% 11% 0.61
III average (50%) 100% 0.65 0.5 25% 5% 0.70
2.2.5. Databases
Site specific climatic profiles appropriately addressing the case study areas were introduced in PestLCI 2.0:
two sets of 30 years average 1971-2000 and 1981-2010 for the Beaucouzé Station, and for the five stations of the
Middle Loire Valley, being the closest to the observed parcels, data for 3 years of production i.e. October n -
September n+1, for 2009-2010 to 2011-2012, a set of averaged months over the 3 years is also available. Climat-
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
ic data were provided by Météo France. The soils of the modelled parcels were characterized through measured
data and observations, in accordance with the PestLCI 2.0 data requirements, and entered in PestLCI 2.0.
2.3. Case study
2.3.1. Choice and main characteristics
Two quite different technical management routes (TMRs) of Chenin Blanc cultivar in the middle Loire Val-
ley (France), were observed during the production year 2010-2011. The prevailing Loire Valley climate is tem-
perate maritime according to the classification used in PestLCI 2.0. This study is focused on vineyard TMRs de-
signed for Protected Denomination of Origin (PDO) dry wine production from this cultivar. The case studies
were chosen within the 5 types of vineyard TMRs resulting from a detailed survey and analysis of existing vine-
yard management practices in the area (Renaud-Gentié et al. 2014): 1 “systematic synthetic chemical use and
limited handwork”,2 “moderate chemical use”, 3 “minimum synthetic treatments and interventions”,4 “moderate
organic”, and 5 “intensive organic”(organic with many interventions and treatments).
The two cases presented here: TMRs 1 and 3 are real situations representing respectively type 1 and 3. Or-
ganic TMRs are not studied here because their treatment programs include mainly or exclusively inorganic pes-
ticide ai.s. No surface water body lies at less than 100m of the parcels. The 2 plots are not drained nor irrigated.
In comparison to the 30 year average (1981-2010) of the area, the climate of the focus year is: i) a little
warmer (0.2°C above the annual average) with a warmer spring but a cooler summer, ii) much drier especially
during the growth season of the vine.
TMR 1’s plot climate is described by Blaison-Gohier weather station, is on the soil “UTB131” according to
the regional cartography of vineyard soils, with 5% slope. TMR 3’s plot climate is described by Fontaine-Guérin
station. Its slope is 3%, and soil is “UTB35”. Field width and length were put at 100m for the two plots. Vine-
yard pest management programs, spraying equipment and pesticide ai.s used by the growers, as well as canopy
development stage and grass cover density and extent are described in table 2.
Table 2: Characteristics of pesticide ai.s applications for 2011 on TMRs 1 and 3. In grey: the inorganic and partially inor-
ganic pesticide ai.s that were not included in this study.*: new entries in PestLCI 2.0. ▪ needed a new CF in USEToxTM
pesticide ai.s Application
rate
Crop type + development
stage
Month of
application Application method
width treated
at a time
TMR1
*Amitrole 0.79 Grass I - all phases april sheltered boom 1.85
Aclonifen 0.31 Grass I - all phases april sheltered boom 1.85
Sulfur 5.89 *Vines II - h80% grass may tunnel sprayer 1.85
Folpet 0.74 *Vines II - h80% grass may tunnel sprayer 1.85
Fosetyl-Aluminium 1.47 *Vines II - h80% grass may tunnel sprayer 1.85
*Fluopicolide 0.12 *Vines II - h80% grass may airblast sprayer 7.4
Fosetyl-Aluminium 1.75 *Vines II - h80% grass may airblast sprayer 7.4
*▪Proquinazid Technique 0.05 *Vines II - h80% grass may airblast sprayer 7.4
*▪Tétraconazole 0.03 *Vines III - a80% grass june airblast sprayer 7.4
*Indoxacarbe 0.04 *Vines III - a80% grass june airblast sprayer 7.4
copper oxychloride 0.73 *Vines III - a80% grass july airblast sprayer 7.4
copper sulfate 0.18 *Vines III - a80% grass july airblast sprayer 7.4
*▪Cymoxanil 0.12 *Vines III - a80% grass july airblast sprayer 7.4
Mancozèbe 0.40 *Vines III - a80% grass july airblast sprayer 7.4
TMR3
Glyphosate 0.54 Grass I - all phases march sheltered boom 1.95
*Amitrole 0.92 Grass I - all phases march sheltered boom 1.95
ammonium thiocyanate 0.86 Grass I - all phases march sheltered boom 1.95
*▪Flazasulfuron 0.02 Grass I - all phases march sheltered boom 1.95
Glyphosate 0.09 Grass I - all phases may sheltered boom 1.95
▪Trifloxystrobin 0.06 *Vines II - a50% grass may pneumatic sprayer side by side 7.8
▪Trifloxystrobin 0.06 *Vines III - a50% grass june pneumatic sprayer side by side 7.8
Diméthomorph 0.18 *Vines III - a50% grass june pneumatic sprayer side by side 7.8
Mancozèbe 1.20 *Vines III - a50% grass june pneumatic sprayer side by side 7.8
*Difénoconazole 0.03 *Vines III - a50% grass july pneumatic sprayer side by side 7.8
*▪Meptyldinocap 0.21 *Vines III - a50% grass july pneumatic sprayer side by side 7.8
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
2.3.2. Calculating characterization factors in USEtox™
Characterization factors (CF) are needed to quantify the potential environmental impact from emissions of
the life cycles of products and systems in LCA. CFs are substance and compartment specific and CFs for each of
the emitted ai.s of the case study for compartments air, surface water and agricultural soil4 are thus needed. We
choose USEtox™ as characterization model since it was developed as a scientific consensus model, to represent
the best application practice for characterization of toxic impacts of chemicals in LCA (Hauschild et al. 2008)
and since its database (v. 1.01) covers ~2500 chemicals with calculated characterization factors for freshwater
ecotoxicity (Rosenbaum et al. 2008). Some of the ai.s were not covered by the database and we therefore applied
the USEtox™ model to calculate them. USEtox™ requires a number of chemical and physical properties in ad-
dition to toxicity data. Primary and secondary sources for these data are available upon demand to the authors.
3. Results
3.1. Emissions and Freshwater ecotoxicity on the case study
Pesticide ai.s emissions were calculated by PestLCI 2.0 for each organic substance application. They are pre-
sented as fraction of applied pesticide ai.s (fig. 1 and 2) and as dose of pesticide ai.s (fig. 3 and 4).
Figures 1 and 2: fraction of applied pesticide ai.s emitted in the 4 compartments: fresh and ground water, nearby
agricultural soil and air (left axis) and the resulting freshwater ecotoxicity (right axis, scales adapted to results)
The total emission fractions do not exceed 0.35, and are lower than 0.15 for most of the pesticides. The emis-
sions loads are dominated by air emissions in figure 1, followed by ground water emissions. In figure 2,
groundwater emissions dominate, for the 4 herbicides. Emissions to nearby agricultural soils are negligible, and
thus don’t appear on the charts. The absence or quasi-absence of freshwater emissions are due to the absence of
water body around the parcels.
The higher total emission fractions relate primarily to Tetraconazole (fungicide of the Triazole group), Cy-
moxanil (fungicide of Cyanoacetamide Oxime group) and Mefenoxam (fungicide of the Phenylamide group).
These fungicides are emission load wise followed by two herbicides: Aclonifen (Diphenyl ether group) and
Amitrole (Triazole group) and another fungicide: Folpet (Phtalimides group).
In some cases, the same substance is applied under two different conditions. Amitrole, for example, is applied
with the same type of boom (herbicide sheltered boom) and on the same canopy (grass). However, the emissions
of this compound vary according to the conditions of application (month of application and soil vary). An appli-
cation of Amitrole in April (figure 1) compared to an application of Amitrole in March (figure 2) shows lower
emissions to air for the early application, while emissions to groundwater are higher for the latest application.
4 USEtox contains no ground water compartment. Ecotoxicological impacts in freshwater from chemical emissions to groundwater are
considered negligible and thus not further considered in this study.
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
Figure 3 and 4: Quantity of applied pesticide ai.s emitted in the 4 compartments and freshwater ecotoxicity (Im-
pacts Ecotox (same values as fig 1 and 2) scale was fixed here to a maximum of 100 to see smaller differences)
When expressing the emissions in kg/ha, see figures 3 and 4, the quantity emitted to the ecosphere per appli-
cation is not higher than 0.12 kg/ha. Amitrole dominates the emissions air in both scenarios (fig. 3 and 4), fol-
lowed by Folpet and Aclonifen.
Freshwater ecotoxicity (FwEtox) was calculated applying USEtox™ characterization factors. Figures 1 and 2
reveal considerable differences for the different application times. In figure 1, FwEtox is dominated by Folpet
(8000 PAF∙m3∙day), Aclonifen (500 PAF∙m3∙day), Fluopicolide (80 PAF∙m3∙day) and Cymoxanil (40
PAF∙m3∙day). The high impact from Folpet emissions is primarily caused by emissions to air and can be ex-
plained from a combination of the substance’s relatively low vapor pressure (meaning that a substantial amount
of emissions of Folpet to air ends up in the freshwater compartment) and its relatively high toxicity to freshwater
organisms. TMR3 shows a much lower total FwEcotox (58.5 PAF∙m3∙day) than TMR1 (9132 PAF∙m3∙day).
3.2. Comparison with static emission quantification approaches
The Ecoinvent approach applied for pesticides is to consider that 100% is emitted to the soil (Nemecek and
Schnetzer 2011), (Neto et al. 2012) in their LCA of Portuguese wine Vinho Verde propose a static partition as
with 75 % of pesticides emitted to soil and 25% to the air. Table 2 compares results between the 3 approaches.
Table 3: Comparison of pesticides emissions calculated by PestLCI 2.0, Ecoinvent and Neto et al. (2012) ap-
proaches and FwEtox calculated with USEToxTM characterization factors
Emissions average fraction emitted standard deviation on fractions
PestLCI 2.0-emissions to air 7.39E-02 9.30E-02
PestLCI 2.0-emissions to surface water 0 0
PestLCI 2.0-emissions to nearby agri. soil 8.06E-05 1.19E-04
PestLCI 2.0-emissions to ground water 1.32E-02 1.68E-02
PestLCI 2.0-total emissions 8.72E-02 9.21E-02
100% emitted to soil (Ecoinvent) 1 0
75% emitted to soil (Neto etal 2012) 0.75E-01 0
25% emitted to air (Neto etal 2012) 0.25E-01 0
Impacts average impact standard deviation on impacts
Impacts FwEcotox PestLCI 2.0 (PAF∙m3∙day) 4.84E+02 1.94E+03
impacts FwEcotox Ecoinvent (PAF∙m3∙day) 1.52E+04 6.46E+04
impacts FwEcotox Neto et al. (PAF∙m3∙day) 1.26E+04 5.35E+04
In the present cases, the average of total emission fraction modelled with PestLCI 2.0 is more than 31 times
lower than the total soil emission fraction estimated by Ecoinvent approach. The average PestLCI 2.0 modelled
emission fraction to air is 3.4 times lower than the total emission fraction to air estimated by Neto et al. (2012)
approach. This leads to differences in FwEtox estimates: 31 times lower with PestLCI model than Ecoinvent ap-
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
proach, and 26 times lower with PestLCI model than with Neto et al. (2012) approach. Very high standard devia-
tions must be noticed on FwEtox, due to high differences in the ai.s’ FwEtox potential.
4. Discussion
While having been intended mainly for arable crops, PestLCI 2.0 inventory model, due to its rather flexible
framework, has here been adapted for viticulture without compromising the model framework. The calculation
of toxicological impact potentials, beyond emission load quantification, was further complicated by ai.s not cov-
ered by the USEToxTM database, thus necessitating the calculation of CFs using the USEToxTM mode.
The case study results show that emission fractions vary in a large extent due to the ai.s properties and parcel
and application conditions. High emissions fractions for most of the ai.s are compensated by very low applica-
tion doses (cymoxanil, teraconazole) leading to moderate emissions (dose) but the ecotoxicological profile of the
ai.s then has high importance on the final FwEtox results as shown by the FwEtox of Folpet and Aclonifen. Mul-
tiple factors differentiate the two case vineyards TMR1 and 3. The main factors are considered to be soil, sprayer
equipment and type of pesticides applied. TMR1 shows higher emissions fractions than TMR3, however the to-
tal emission load is lower because of the low doses applied for some substances. Characterization through USE-
ToxTM reveals different results for the two cases, mainly due to the high ecotoxicity potential of Folpet and to a
lesser extent, of Aclonifen, even if both ai.s are applied via sheltered boom and tunnel sprayer (both limiting
wind drift) (TMR1).
Inorganic pesticide ai.s, which are not modelled here, were also applied to the case vineyards: five applica-
tions for TMR1 including copper based ai.s and one application for TMR3 (Table2). The copper based pesticide
ai.s are expected to further increase the FwEtox of TMR1 if included (Mackie et al. 2012; Vázquez-Rowe et al.
2012). A comparison of the present TMRs FwEtox profiles with the results obtained by (Vázquez-Rowe et al.
2012) with PestLCI 1 in Galician vineyards shows very good performance of the present TMR. Hence, TMR1’s
FwEtox is half of the lowest FwEtox mentioned by this author (Copper impacts removed) while TMR3’s
FwEtox is 400 times lower as this author’s lower result.
Huge differences were found between the two static approaches (Ecoinvent and Neto et al. 2012) and Pes-
tLCI 2.0 based emission quantification. Accounting for the sole emissions that cross the parcel borders is a first
element lowering the quantity of emitted pesticides as modelled by PestLCI 2.0, compared to the other ap-
proaches tested. However that is not the only cause of lower emissions and FwEtox; considering, for these emis-
sions calculations, processes of evaporation, runoff and leaching, canopy influence, including the actual proper-
ties of the pesticide ai.s applied, of soils and sprayers allows for a more accurate adjustment of estimates to the
real phenomenons. Degradation of pesticide ai.s and their uptake by the plants are actual processes that aren’t
considered in the static approaches tested, but accounted for in PestLCI 2.0 (Dijkman et al. 2012). The uptake of
pesticides in crops including edible parts, is in PestLCI only considered as a fate/removal process and hence ex-
port from the field of pesticides via crops and crop residues is not considered nor quantified. Plant pesticide up-
take processes are complex systems of coupled processes demanding dedicated models such as DynamiCROP
(see e.g. (Fantke et al. 2011)).
In organic viticulture, sulfur and copper are the only means available to manage respectively powdery and
downy mildew, and represent important quantities of applied pesticides. Sulfur and copper are not available in
PestLCI 2.0 as the model is designed for modelling of organic pesticide ai. emissions. Thus, a comparison be-
tween conventional and organic viticulture or inclusion of organic managed cases in a study cannot be dealt with
solely through PestLCI 2.0; a model similar to PestLCI is needed for inorganic pesticides. Considering the com-
plex chemistry of inorganic chemicals, such a model will however have to deal with multiple species nature of
inorganic chemicals in order to appropriately reflect the behaviour of these chemicals in soils.
PestLCI 2.0 could be improved by further developments of the accounting of airborne drift, which can be
considerable (Jensen and Olesen 2014) but the complexity of the phenomena (Gil et al. 2008) and the lack of
(generic) data are considered major obstacles for this improvement. More or less for the same reasons, pesticide
metabolites are not accounted for in the present version of PestLCI 2.0. Accounting for application parameters as
sprayers’ speed, droplets size, temperature, relative humidity would be ideal for further refinement of the model-
ling of the spray mixture behaviour and fate, but these parameters are too difficult to obtain from the growers,
and would further entail an even more complicated inventory.
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
High percentages of stones can be found in many vineyard soils; and modify water flow in the soil. These as-
pects could not be included in the present customization of PestLCI 2.0. However we recommend improvement
of the way soil texture affects macropore transport in PestLCI 2.0 as an important issue to be considered in the
coming PestLCI versions.
PestLCI 2.0, calculations must be carried out for each pesticide ai.s and application and thus remain a tedious
task. Ideally, a possible connection with an external Excel database could potentially facilitate the automation
(batch) of calculation and further pave the road for more detailed sensitivity analyses.
5. Conclusion
The PestLCI 2.0 customized version for viticulture, presented in the paper at hand, facilitates the calculations
of emission loads for vertically trained vineyards with a wide range of sprayers and further provides a considera-
ble PestLCI pesticide database update (even if non-exhaustive), of viticulture specific pesticides ai.s, and is
complemented by the corresponding UseTOXTM FwEtox characterization factors. A range of rather important
specificities of viticulture that were not included in PestLCI 2.0. has been added like: i) specific sprayers drift
characteristics, ii) the presence of a secondary canopy (cover-crop), iii) specific data like specific pesticide ai.s in
the PestLCI database.
Application on two different case studies shows that emissions differ from a pesticide ai.s application to the
other due to different emissions fractions related to environmental conditions and ai.s properties.
Huge differences were found between the two static approaches (Ecoinvent and Neto et al. 2012) and Pes-
tLCI 2.0 based emission quantification and FwEtox, The static approaches over-evaluate emissions and hence
FwEtox from 2 to 31 times compared to PestLCI 2.0, due to different technosphere/ecosphere boundaries and
accounting of degradation and plant uptake phenomena by PestLCI 2.0.
Some of the new PestLCI model parameters can also be used for other perennial or bush crops as long as
equipment, shape of the canopy and pesticides a.i.s stay in the range of available options.
Aknowledgements: The authors thank Interloire and Region pays de la Loire for funding, the winegrowers for
their time and data, Miss D. Boudiaf for her help in data collection, Mrs F. Jourjon and Mrs E. Mehinagic for
their core reading and advices.
6. References
Ambiaud E (2012a) Moins de désherbants dans les vignes. vol Oct 2012. Agreste, statistique agricole,
Ambiaud E (2012b) Pratiques phytosanitaires dans la viticulture en 2010, SSP - Bureau des statistiques végétales
et animales. vol Oct 2012.
Ardente F, Beccali G, Cellura M, Marvuglia A (2006) POEMS: A Case Study of an Italian Wine-Producing
Firm. Environmental Management 38 (3):350-364. doi:10.1007/s00267-005-0103-8
Birkved M, Hauschild MZ (2006) PestLCI--A model for estimating field emissions of pesticides in agricultural
LCA. Ecological Modelling 198 (3-4):433-451. doi:10.1016/j.ecolmodel.2006.05.035
Bosco S, Di Bene C, Galli M, Remorini D, Massai R, Bonari E (2011) Greenhouse gas emissions in the
agricultural phase of wine production in the Maremma rural district in Tuscany, Italy. Italian Journal of
Agronomy 6 (2):93-100. doi:10.4081/ija.2011.e15
Brausch J, Smith P (2007) Toxicity of Three Polyethoxylated Tallowamine Surfactant Formulations to
Laboratory and Field Collected Fairy Shrimp, Thamnocephalus platyurus. Arch Environ Contam Toxicol 52
(2):217-221. doi:10.1007/s00244-006-0151-y
Celette F, Findeling A, Gary C (2009) Competition for nitrogen in an unfertilized intercropping system: The case
of an association of grapevine and grass cover in a Mediterranean climate. European Journal of Agronomy 30
(1):41-51. doi:http://dx.doi.org/10.1016/j.eja.2008.07.003
Codis S (2014) discussion about interception of spray mixture by vineyard and grass cover at different stage of
vine growth, and drift measurements on vineyard sprayers.
Codis S, Bos C, Laurent S (2011) Réduction de la dérive, 8 matériels testés sur vigne. Phytoma 640 (Jan
2011):1-5
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
Dijkman T, Birkved M, Hauschild M (2012) PestLCI 2.0: a second generation model for estimating emissions of
pesticides from arable land in LCA. The International Journal of Life Cycle Assessment:1-14.
doi:10.1007/s11367-012-0439-2
Fantke P, Charles R, Alencastro LFd, Friedrich R, Jolliet O (2011) Plant uptake of pesticides and human health:
Dynamic modeling of residues in wheat and ingestion intake. Chemosphere 85 (10):1639-1647.
doi:http://dx.doi.org/10.1016/j.chemosphere.2011.08.030
Ganzelmeier H (2000) Drift studies and drift reducing sprayers -a german approach-. Paper presented at the 2000
ASAE annual international meeting, Milwaukee, Wisconsin, July 9-12 2000
Gil Y, Sinfort C, Brunet Y, Polveche V, Bonicelli B (2007) Atmospheric loss of pesticides above an artificial
vineyard during air-assisted spraying. Atmospheric Environment 41 (14):2945-2957.
doi:http://dx.doi.org/10.1016/j.atmosenv.2006.12.019
Gil Y, Sinfort C, Guillaume S, Brunet Y, Palagos B (2008) Influence of micrometeorological factors on pesticide
loss to the air during vine spraying: Data analysis with statistical and fuzzy inference models. Biosystems
Engineering 100 (2):184-197. doi:http://dx.doi.org/10.1016/j.biosystemseng.2008.03.009
Jensen PK, Olesen MH (2014) Spray mass balance in pesticide application: A review. Crop Protection 61 (0):23-
31. doi:http://dx.doi.org/10.1016/j.cropro.2014.03.006
Linders J, Mensink H, Stephenson G, Wauchope D, Racke K (2000) Foliar Interception and Retention Values
after Pesticide Application. A Proposal for Standardized Values for Environmental Risk Assessment
(Technical Report). Pure and Applied Chemistry 72 (11):2199-2218. doi:10.1351/pac200072112199
e-phy, Le catalogue des produits phytopharmaceutiques et de leurs usages des matières fertilisantes et des
supports de culture homologués en France (2013) Ministère de l'Agriculture, de l'Agroalimentaire et de la
Forêt, Organisation Nationale pour la Protection des Végétaux. http://e-phy.agriculture.gouv.fr/. Accessed
october-december 2013 and january 2014
Mackie KA, Müller T, Kandeler E (2012) Remediation of copper in vineyards – A mini review. Environmental
Pollution 167 (0):16-26. doi:10.1016/j.envpol.2012.03.023
Mézière D, Gary C, Barbier JM, Bernos L, Clément C, Constant N, Delière L, Forget D, Grosman J, Molot B,
Pio P, Sauvage D, Sentenac G (2009) ECOPHYTO R&D. Vers des systèmes de cultures économes en
produits phytosanitaires. Volet 1 : Analyse comparative de différents systèmes en viticulture (Tome III).
INRA,
Nemecek T, Schnetzer J (2011) Methods of assessment of direct field emissions for LCIs of agricultural
production systems, Data v3.0 (2012).
Neto B, Dias AC, Machado M (2012) Life cycle assessment of the supply chain of a Portuguese wine: from
viticulture to distribution. The International Journal of Life Cycle Assessment:1-13. doi:10.1007/s11367-012-
0518-4
Pattara C, Raggi A, Cichelli A (2012) Life Cycle Assessment and Carbon Footprint in the Wine Supply-Chain.
Environmental Management 49 (6):1247-1258. doi:10.1007/s00267-012-9844-3
Point E, Tyedmers P, Naugler C (2012) Life cycle environmental impacts of wine production and consumption
in Nova Scotia, Canada. Journal of Cleaner Production 27 (0):11-20. doi:10.1016/j.jclepro.2011.12.035
Renaud-Gentié C, Burgos S, Benoît M (2014) Choosing the most representative technical management routes
within diverse management practices: Application to vineyards in the Loire Valley for environmental and
quality assessment. European Journal of Agronomy 56 (0):19-36.
doi:http://dx.doi.org/10.1016/j.eja.2014.03.002
Renaud-Gentié C, Burgos S, Benoît M, Jourjon F (2013) Evaluation environnementale d’itinéraires viticoles
représentant une diversité régionale par analyse de cycle de vie/Environmental evaluation of vineyard
Technical Management Routes representing a regional diversity using the Life Cycle Assessment. Progrès
Agricole et Viticole 130eme année (10-2013):11-17
Rosenbaum R, Bachmann T, Gold L, Huijbregts MJ, Jolliet O, Juraske R, Koehler A, Larsen H, MacLeod M,
Margni M, McKone T, Payet J, Schuhmacher M, Meent D, Hauschild M (2008) USEtox—the UNEP-
SETAC toxicity model: recommended characterisation factors for human toxicity and freshwater ecotoxicity
in life cycle impact assessment. The International Journal of Life Cycle Assessment 13 (7):532-546.
doi:10.1007/s11367-008-0038-4
Chemspider, search and share chemistry (2013) Royal Society of Chemistry. Accessed october-december 2013
and january 2014
9th International Conference LCA of Food San Francisco, USA 8-10 October 2014
Sinfort C (2014) discussion about interception of spray mixture by vineyard and grass cover at different stage of
vine growth.
Sinfort C, Cotteux E, Bonicelli B, Ruelle B (2009) Une méthodologie pour évaluer les pertes de pesticides vers
l'environnement pendant les pulvérisations viticoles. Paper presented at the STIC & Environnement, Calais,
France, 2009
The Bio-Pesticide DataBase (BPDB) developed by the Agriculture & Environment Research Unit (AERU),
(2012) University-of-Hertfordshire 2012 Accessed october-december 2013-january 2014
The Pesticide Properties DataBase (PPDB) developed by the Agriculture & Environment Research Unit
(AERU), (2013) University-of-Hertfordshire 2006 - 2013.
US-Environmental-protection-Agency (2012) EPI SuiteTM v4.11. US Environmental protection Agency,
TOXNET - Databases on toxicology, hazardous chemicals, environmental health, and toxic releases. (2013) US
national library of medicine. Accessed october-december 2013 and january 2014
van Zelm R, Larrey-Lassalle P, Roux P (2014) Bridging the gap between life cycle inventory and impact
assessment for toxicological assessments of pesticides used in crop production. Chemosphere 100 (0):175-
181. doi:http://dx.doi.org/10.1016/j.chemosphere.2013.11.037
Vázquez-Rowe I, Villanueva-Rey P, Moreira MT, Feijoo G (2012) Environmental analysis of Ribeiro wine from
a timeline perspective: Harvest year matters when reporting environmental impacts. Journal of Environmental
Management 98 (0):73-83. doi:10.1016/j.jenvman.2011.12.009
Villanueva-Rey P, Vázquez-Rowe I, Moreira MT, Feijoo G (2014) Comparative life cycle assessment in the
wine sector: biodynamic vs. conventional viticulture activities in NW Spain. Journal of Cleaner Production
65 (0):330-341. doi:http://dx.doi.org/10.1016/j.jclepro.2013.08.026